EP0540762B1 - Procede pour detecter un vecteur de mouvement et appareil a cet effet, et systeme pour traiter un signal video a l'aide de cet appareil - Google Patents

Procede pour detecter un vecteur de mouvement et appareil a cet effet, et systeme pour traiter un signal video a l'aide de cet appareil Download PDF

Info

Publication number
EP0540762B1
EP0540762B1 EP92910679A EP92910679A EP0540762B1 EP 0540762 B1 EP0540762 B1 EP 0540762B1 EP 92910679 A EP92910679 A EP 92910679A EP 92910679 A EP92910679 A EP 92910679A EP 0540762 B1 EP0540762 B1 EP 0540762B1
Authority
EP
European Patent Office
Prior art keywords
detecting
motion vector
vector
picture
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
EP92910679A
Other languages
German (de)
English (en)
Other versions
EP0540762A4 (en
EP0540762A1 (fr
Inventor
Hiroshi Nippon Hoso Kyokai Hoso Hirabayashi
Yuji Nippon Hoso Kyokai Hoso Gijutsu Nojiri
Yasuaki Nippon Hoso Kyokai Hoso Kanatsugu
Shoichi Nippon Hoso Kyokai Hoso Gijutsu Suzuki
Hajime Nippon Hoso Kyokai Hoso Sonehara
Junji Nippon Hoso Kyikai Hoso Gijutsu Kumada
Iwao Nippon Hoso Kyokai Hoso Gijutsu Obata
Kenji Nippon Hoso Kyokai Hoso Senta Nakajima
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Japan Broadcasting Corp
Original Assignee
Nippon Hoso Kyokai NHK
Japan Broadcasting Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from JP14662491A external-priority patent/JP2898787B2/ja
Priority claimed from JP3146625A external-priority patent/JP3067275B2/ja
Priority claimed from JP16625491A external-priority patent/JP2989325B2/ja
Priority claimed from JP20117091A external-priority patent/JP2898798B2/ja
Application filed by Nippon Hoso Kyokai NHK, Japan Broadcasting Corp filed Critical Nippon Hoso Kyokai NHK
Priority to EP97110945A priority Critical patent/EP0806866B1/fr
Publication of EP0540762A1 publication Critical patent/EP0540762A1/fr
Publication of EP0540762A4 publication Critical patent/EP0540762A4/en
Application granted granted Critical
Publication of EP0540762B1 publication Critical patent/EP0540762B1/fr
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51—Motion estimation or motion compensation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G06T7/269—Analysis of motion using gradient-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N5/00—Details of television systems
    • H04N5/14—Picture signal circuitry for video frequency region
    • H04N5/144—Movement detection
    • H04N5/145—Movement estimation

Definitions

  • the present invention relates to a method of detecting motion of a picture on an image or picture plane as a motion vector from a picture signal by using a gradient method, an apparatus for carrying out the method and a picture signal processing system for performing picture signal processings such as high-efficiency picture signal encoding, frame rate conversion and others by making use of the motion vector as detected.
  • the gradient method has been described on the basis of the picture signal.
  • a method of determining the motion vector on the basis of pictures as generated on a display screen or a picture plane will be elucidated by reference to Fig. 7.
  • the picture plane is divided into a plurality of blocks with n and m blocks in the x- and y-directions, respectively.
  • the display picture plane for display is divided into 16 blocks with 4 blocks in each of the x- and y-directions.
  • An object D0 displayed within a block 200-33 in a preceding field is displayed as D1 within a block 200-22 in the current field.
  • the motion of the object at this time can be represented by a motion vector V.
  • a block gradient method for detecting this motion vector V, there is generally adopted a block gradient method.
  • a picture plane is divided into a plurality of blocks, wherein a motion vector to a given one of the blocks is determined by using inter-field or inter-frame signal differences at a plurality of pixels i belonging to each block.
  • the motion vector V can be expressed as follows: Let's consider an x-directional component and a y-directional component of a motion vector on an assumption that K pixels exist in each block.
  • the x-component V x and the y-component V y of the motion vector V can be expressed as follows: where symbol ⁇ indicates summation for all the pixels within a given block.
  • the motion vector V of a television picture is to be determined by the gradient method
  • a spatial band limitation is applied to the picture signal so that the picture signal has a same gradient at one and the same place between different fields or frames with a view to increasing detection accuracy.
  • those signal components which are modulated and folded back by an interlacing carrier are diminished through a sequential scanning.
  • the circuit for the sequential scanning performs generally an intra-frame processing for a still picture portion while performing an intra-field processing for a motion picture portion.
  • Fig. 4A there are illustrated light signals impinging into a TV camera at a field interval.
  • a solid line, a broken line, a single-dot line, a double-dot line and a triple-dot line represent signals at an n-th field, (n + 1)-th field, (n + 2)-th field, ((n + 3)-th field and an (n + 4)-th field, respectively.
  • a rising portion of the edge moves toward the right.
  • an output signal therefrom will be of such waveform as illustrated in Fig.
  • the denominator of the expressions (1) and (2) i.e. the sum of absolute gradient values, assumes a value approximating to zero in a region where the gradient is small, as is obvious from the expressions (1) and (2).
  • the value of the detected motion vector becomes significantly greater than the true or intrinsic motion vector value, thus presenting a problem that remarkable error is involved.
  • detection of a scene change in a television picture is carried out in such manner that inter-frame or inter-field differences in the picture signals is determined over a plurality of regions on a picture plane, wherein when the difference exceeds a certain threshold value, motion of the picture is decided and when the regions for which the motion of the picture is decided exceeds a certain proportion of one scene, the scene change of the picture is detected.
  • the scene change detecting method is disadvantageous in that upon appearance of many picture signals representing motion in a scene due to planning and tilt, the scene change is detected erroneously.
  • the present invention teaches an apparatus, a system and a method as described in claims 1, 10 and 17, respectively.
  • a motion vector detecting apparatus comprising:
  • a picture signal processing system comprising:
  • a method of detecting a motion vector from picture signals comprising the steps of:
  • the detected vector exhibit significantly large values when compared with the intrinsic vector value in the region of the picture signal where the gradient is small, giving rise to erroneous vector detection.
  • the detected vector can assume a value approximating to zero in the region where the gradient is small without exerting any appreciable influence to the region where gradient is large.
  • embodiments of the present invention may be capable of detecting with improved accuracy a motion vector representing a motion of an object moved on a picture plane.
  • the input picture signal may be added with an electrically generated after-image or residual image so as to be temporally equivalent to a picture signal undergone an LPF (low-pass filter) processing, as a result of which a region of the picture signal in which the gradient remains same between the fields or frames is enlarged to improve accuracy in detecting the motion vector by the gradient method.
  • LPF low-pass filter
  • reliability of the detected motion vector of a picture need not be determined on the basis of an inter-frame difference between two corresponding picture signals but is preferably determined on the basis of a value obtained by standardizing a distance on an N-dimensional coefficient space realized through an orthogonal expansion of one-dimensional or two-dimensional functions representing the picture.
  • reliability of decision as to appropriateness of the motion vector as determined can be enhanced over the prior art method.
  • the degree of similarity between two picture undergoing the comparison can be displayed quantitatively.
  • detection of scene change is preferably effected by adopting a method of determining the motion vector of a picture by a recursive block gradient method which is based on the use of the initial displacement vector and by evaluating the number of sub-blocks of the picture plane where the magnitude of the displacement vector determined by a first or initial gradient method exceeds a first threshold value.
  • the motion vector detecting apparatus includes a temporal band limiter section 2, a field memory group section 3, a spatial band limiter section 4 and a detecting section 5.
  • the band limiter section 2 destined to apply a temporal band limitation to an unprocessed or reference picture signal as inputted includes four field memories 2-1 to 2-4 for storing sequentially picture signals each of one field, respectively, five scalers 2-11 to 2-15 for weighting the picture signals outputted from the field memories 2-1 to 2-4, respectively, and a summing circuit 2-20 for adding together the signals weighted by the scalers 2-11 to 2-15.
  • the field memories 2-1 to 2-3 transfer the picture signals stored therein to the adjacent field memories 2-2 to 2-4, respectively, on a field-by-field basis.
  • the field memory 2-1 stores the input picture signal, while the picture signal stored in the field memory 2-4 is used for other purpose.
  • the field memory group section 3 for storing the output signals from the limiter section 2 as temporally filtered picture signals includes a current field memory 14-1 for storing the temporally filtered picture signal of the current field and a preceding field memory 14-2 for storing the temporally filtered picture signal of the preceding field.
  • the signal stored in the current field memory 14-1 in the current field is stored in the preceding field memory 14-2 in the succeeding field.
  • the spatial band limiter section 4 includes spatial filters 16-1 and 16-2 for spatially filtering the picture signals stored in the current field memory 14-1 and the preceding field memory 14-2, respectively.
  • the detecting section 5 for detecting the motion vector includes a motion vector memory 18 for storing the motion vectors detected already, a motion vector selector 20 for selecting one of the motion vectors stored in the memory 18 as a motion vector V 0 in a accordance with the output signals from the filters 16-1 and 16-2, a motion vector detecting stage 22a for detecting a motion vector V 1 from the output signals of the filters 16-1 and 16-2 by using the motion vector V 0 from the selector 20 and an adder 24a for generating a motion vector V from the motion vectors V 0 and V 1 .
  • the temporal band limiter section 2 in a stage preceding to the sections 3, 4 and 5 operative on the basis of the gradient method.
  • the aim of this limiter section 2 is to generate electrically an after-image (residual image) effect of a TV camera for the purpose of enlarging an overlap region in the picture signals of different fields or frames which region has ideally a same gradient.
  • the limiter section 2 realized in the form of a temporal filter having five taps by using four field memories 2-1 to 2-4.
  • Coefficients k1 to k5 of the scalers 2-11 to 2-15 represent tap coefficients of the five-tap filter.
  • the gradient method it is required for detecting accurately the motion vector that the gradient of the picture signal subjected to the detection remains coincident or throughout the fields.
  • the gradient method can not be used.
  • the rising portion of the edge of the picture signal has a gradient owing to a storage effect, as is shown in Fig. 4B.
  • the gradient of the rising edge portion will be more gentle due to presence of the after-image effect as well, meaning that the region in which the gradient method can be used increases.
  • the motion vector is detected by making use of the storage effect and the after-image effect of the television camera.
  • the after-image effect of the television camera With the after-image effect of the television camera, a part of the picture signal of the preceding field is added to that of the current field. However, the picture signal of the current field can never affect that of the preceding field. In a television camera where a CCD (Change-Coupled Device) is used, the after-image effect is insignificant. Moreover, when a shutter is inserted, the storage effect also decreases. In that cases, characteristics of the temporal band limiter section 2 or the temporal filter provided according to the invention is adjusted in dependence on magnitude of the storage or after-image effect of the television camera as used to thereby generate electrically the storage or after-image effect.
  • CCD Change-Coupled Device
  • the picture signal outputted from the summing unit 2-20 is stored in the current field memory 14-1 of the section 3.
  • the picture signal stored in the current memory 141 is transferred to the preceding field memory 14-2 in precedence to the storage of the picture signal outputted from the summing unit 2-20.
  • the picture signals stored in the memories 14-1 and 14-2 are supplied to the spatial filters 16-1 and 16-2, respectively, of the section 4 to undergo spatial band limitation.
  • the band-limited picture signals are supplied to a motion vector selector 20 of the and the motion vector detecting stage 22a of the section 5.
  • the display picture plane is divided into a plurality of blocks each having a plurality of pixels.
  • the selector 20 selects an initial displacement vector relative to a concerned block of the current field from the motion vector memory 18.
  • a predetermined number of candidates for the motion vector are selected from the motion vector of the block B 0 located in the preceding field at the position corresponding to the concern block B of the current field, the motion vectors of the blocks surrounding the block B 0 and the motion vectors detected already for the blocks belonging to the current field.
  • a block of a same size as the block B 0 is set in the preceding field at a position to which the block B 0 of the preceding field has been displaced in correspondence to the motion vector thereof, and then a pattern matching is performed between the set block and the concerned block B.
  • the motion vector for a block B 1 exhibiting the best pattern matching is then selected as the initial displacement vector V 0 which is supplied to the detecting stage 22a and the adder 24a.
  • the detecting stage 22a includes for individual components of the motion vector a subtraction unit 32, an x- or y-differentiation filter 34, a multiplier 36, a summing unit 38, a sign extracting unit 40, an adder 42, an absolute value circuit 44, a summing unit 46, a divider 48 and a dither signal generator 49.
  • V x for example, the current field picture signal from the filter 16-1 is supplied to the differentiation filter 34 and the subtraction unit 32, while the preceding field picture signal from the filter 16-2 is supplied to the subtraction unit 32.
  • the differentiation filter 34 differentiates the current field picture signal to thereby determine gradient values g xi of the current field picture signal at the individual pixels i within the concerned block.
  • the gradient values are then supplied to the adder 42 and the circuit 40.
  • sign of the gradient value is extracted to be outputted to the multiplier 36.
  • the subtraction unit 32 subtracts the current field picture signal from the preceding field picture signal for each of the pixels within the concerned block and outputs differences ⁇ t i resulting from the subtraction to the multiplier 36, which codes the values ⁇ t i on the basis of the sign signal supplied from the circuit 40, the coded values being outputted to the summation unit 38.
  • the summation unit 38 sums the outputs of the multiplier 36 for all the pixels of the concerned block.
  • the adder 42 which is supplied with a minute dither signal from the dither signal generator 49 adds the gradient values g xi with a value S of the dither signal, the result of the addition being outputted to the circuit 44, which then determines the absolute value of the output from the adder 42, the result of which is outputted to the summation unit 46.
  • the summation unit 46 the outputs from the circuit 44 are summed for all the pixels of the concerned block.
  • the divider 48 divides the output of the summation unit 38 by that of the summation unit 46, the result of which is outputted as a value V x or V y of one component of the motion vector V.
  • the detected vector V 1 is supplied to the adder 24a to be added with the initial displacement vector V 0 , as a result of which the motion vector V of the concerned block in the current field can be determined.
  • This motion vector is stored in the motion vector memory 18.
  • V x and V y of the motion vector V can be expressed as follows: where S x and S y represent S , i.e., value which makes appearance owing to the addition of the dither signal on the block basis and assumes different magnitudes in dependence on the types of the pictures.
  • Influence or contribution of the sums S x or S y of the dither signal value is zero or approximates closely to zero in a region where the gradient is of large magnitude. However, as the gradient becomes smaller, the contribution of the sum S x or S y previously increases in the relative sense. More specifically, values of the x-component V x and the y-component V y of the motion vector V given by the expressions (3) and (4), respectively, are insusceptible to the influence due to addition of the dither signal in a region where the gradient is of large magnitude, while the values of these vector components become smaller in the region where the gradient is of small because the values of denominators in the expressions (3) and (4) increases.
  • the method of adding the dither signal value to the gradient is more effective for the detection of the motion vector by the gradient method using the initial displacement vector V 0 .
  • the method in which addition of the dither signal to the gradient signal is not adopted will involve significant error in the detected displacement vector V 1 because the components V x1 and V y1 of the displacement vector V 1 determined from the expression (2) assume significantly large values in a region where the gradient is small.
  • the components V x1 and V y1 are close to zero even in the region of small gradient, thus ensuring that the displacement vector V 1 as detected can assume a value approximating to the initial displacement vector V 0 , whereby possibility of occurrence of significant error can positively be excluded.
  • the temporal band limiter section is not always indispensably required but may be spared so long as the reference picture signal can provide sufficiently high storage and after-image effects.
  • the picture signal may be processed on a frame-by-frame basis instead of the field-by-field basis.
  • the initial displacement vector is supplied to the detecting stage 22a from the selector 20.
  • the motion vector memory 18 and the selector 20 may be spared.
  • the vector V 1 detected in the detecting stage 22a represents the motion vector V.
  • the absolute value is determined after addition of the dither signal.
  • a block B 2 displaced or deviated from the block B 1 by the vector V 1 is now set in the preceding field, and a displacement vector V 2 is determined between the picture signals of the block B 2 and B in accordance with the block gradient method.
  • this processing is sequentially and recursively performed until the displacement vector converges to zero.
  • the processing is terminated after an appropriate number of the repetitions or when magnitude of the displacement vector has decreased below a preset threshold value.
  • the rate of convergence is usually so high that three times as many as execution of the arithmetic operation according to the recursive block gradient method will be sufficient, as can be seen from Fig. 5.
  • a recursive block gradient method which realizes the concept elucidated above.
  • This method can be realized with the structure shown in Fig. 5. It should however be understood that the basic structure to this end is similar to that shown in Fig. 1 except for a difference in that additional motion vector detecting stages 22b and 22c are provided in addition to the motion vector detecting stage 22a in the case of the arrangement shown in Fig. 5.
  • Each of the detecting stages 22b and 22c is implemented in a configuration similar to that of the detecting stage 22a.
  • the detecting stages 22a, 22b and 22c are connected in series to one another, wherein a displacement vector is supplied from the preceding stage to the succeeding one with the initial displacement vector being supplied to the leading stage.
  • the motion vector detected finally is stored in the memory 18.
  • the displacement vector V 1 is supplied to a scene change detecting stage 6 which includes a magnitude decision circuit 52, a counter 54 and a scene change detector 56.
  • the displacement vector V 1 outputted from the detecting stage 22a is supplied to the circuit 52 where magnitude of the displacement vector V 1 is compared with a first predetermined value.
  • the content of the counter 54 is incremented by "1".
  • This processing is executed for the whole field, i.e., for all the blocks of the picture plane.
  • the count value at that time point is outputted to the detector 56, which then makes decision as to whether or not the count value as inputted is greater than a second predetermined value.
  • a second predetermined value When it is decided that the count value is greater than the second predetermined value, this means that occurrence of a scene change is detected, whereby a scene change signal is outputted.
  • the motion vector memory 18 there are stored successively in the motion vector memory 18 the motion vectors detected for a large number of blocks in the prepreceding field, preceding field and/or the current field. These motion vectors are used as the candidate vectors for determining the initial displacement vector of a concerned block which is to be determined. Subsequently, the initial displacement vector V 0 is selected from the above-mentioned candidate vectors by the initial displacement vector selector 20. Succeeding recursive block gradient arithmetic stages 22a, 22b and 22c are each implemented in a configuration similar to the arithmetic stage 22a operating in accordance with the block gradient method described previously.
  • the very motion vector detecting method elucidated above is utilized.
  • the initial displacement vector V 0 is selected at a value approximating to the intrinsic vector value. Consequently, the displacement vector V 1 determined through the initial arithmetic processing based on the gradient method often assumes a smaller value.
  • the scene change differs from the picture in which an object or objects are moving. Accordingly, it is meaningless to determine the motion vector itself. Nevertheless, if it is presumed to determine the motion vector at any rate, there will be selected some vector as the initial displacement vector V 0 through pattern matching, and the displacement vector V 1 will exhibit a tendency to assume a large value.
  • an output signal representing occurrence of a scene change may be produced at the time point when the coefficient value for one field or one frame exceeds a second predetermined value in the scene change detector 56.
  • the second predetermined value may be given in terms of a certain ratio to a total number of the blocks in one picture plane or scene.
  • degree of similarity of two pictures represented by functions f(x, y) and g(x, y) is evaluated on the basis of a distance between two points on a coefficient space determined by coefficients obtained from orthogonal function expansion of the functions f(x, y) and g(x, y) or alternatively by using data obtained by standardizing the above-mentioned distance by distances from the origin of the scalar space or the like factor.
  • a high degree of similarity between the two pictures will eventually show that the reliability of the detected vector is high.
  • the similarity evaluation method can be utilized for checking the reliability of the motion vector.
  • the two points on the N-dimensional space coincide with each other, while the distance between the two pints on the N-dimensional space increase when difference between the two pictures becomes more significant.
  • the distance between the two points as the measure of similarity between the two points. Further, by standardizing the above distance with distances to these points from the origin of the N-dimensional space, dispersions of the similarity from one to another picture can be suppressed.
  • the degree of similarity H can be defined by the following expression. In this way, the degree of similarity between the two pictures or images can be evaluated by taking into account the quantitative components as well.
  • the section 7 includes a ROM for storing orthogonal function series, first to N-th orthogonal coefficient scalers 62-1 to 62-N, summing units 64, 66-1 and 66-2, an adder 67, a divider 68 and an evaluation unit 70.
  • Each of the first to N-th orthogonal coefficient scalers 62-1 to 62-N is supplied with picture signals f(x) and g(x) of two blocks to be checked in respect to the degree of similarity from the current field memory 14-1 and the preceding field memory 14-2, respectively.
  • the coefficient scaler 62-1 includes multipliers 72 and 78, summing units 74 and 78, a subtraction unit 82 and squaring units 76-1, 76-2 and 84.
  • the multipliers 72 and 78 determine products of the input picture signal function f(x) or g(x) and the orthogonal function series ⁇ k (x), and the products as determined are summed by the summing units 74 and 80.
  • the one-dimensional coefficients C k and D k will have been determined by the summing units 74 and 80, respectively.
  • the coefficient C k is supplied to the squaring unit 76-1 and the subtraction unit 82, while the coefficient D k is supplied to the squaring unit 76-2 and the subtraction unit 82.
  • subtraction of the coefficients is effected, i.e., (C k -D k ) is determined, the result of which is outputted to the squaring unit 84.
  • the squaring units 76-1, 76-2 and 84 square values inputted thereto, the results of which are outputted to the summing units 66-1, 66-2 and 64.
  • the summing unit 66-1 and 66-2 determine the distances to the points on the coefficient space corresponding to the coefficients C k - D k from the origin of that space.
  • distance between the spatial points corresponding to the coefficients C k - D k is determined by the summing unit 64.
  • the contents of the summing units 66-1 and 66-2 are added together by the adder 67 and the resulting sum is supplied to the divider 68.
  • the divider 68 divides the output value of the summing unit 68 by the output value from the adder 67, the result H of the division being outputted to the evaluation unit 70, which then evaluates the degree of similarity between the two pictures on the basis of the output value H from the divider 68 and additionally checks the reliability of the motion vector.
  • the value outputted from the divider 68 becomes smaller, while the value increases as the degree of similarity becomes lower. Since this is in opposition to the generally adopted standpoint, reciprocal of the above-mentioned value may be determined and handled as the value representing the similarity or evaluation value.
  • (1-H) may be handled as the evaluation value, as is apparent from the following description.
  • a polynomial Pn,m(x) of degree n is, for example, defined as follows, wherein n represents the degree, m represents the number of pixels within a block, and a symbol " ⁇ " represents powers.
  • the initial displacement vector determined by the pattern matching method is progressively modified so as to reach finally determination of the intrinsic motion vector. Consequently, in the recursive block gradient method, although the vector sum as determined gradually approaches to the intrinsic motion vector so long as the displacement vectors determined in the course of the recursive process remain appropriate, the vector sum may diverge when the vector as determined is deviated from the intrinsic displacement vector. In order to prevent such divergence, it is known that magnitude of the displacement vector determined in the course of the recursive process is limited.
  • the divergence can be prevented by multiplying the displacement vector determined in the recursive process by the reliability representing similarity H determined according to the invention in the form of (1-H). So long as pictures remain same, the numerical value of (1-H) is "1", meaning that the vector as determined can be employed intact, while for the pictures differing from one to another, the numerical value of (1-H) becomes close to zero to suppress the divergence.
  • the vector allocation circuit serves for determining an optimal vector from candidate vectors inclusive of zero vector when a plurality of vectors are determined in a block or from a set of a vector and zero vector when one vector is determined in a block.
  • a conventional method of determining the optimal vector resides in a procedure of calculating difference between two pictures by moving one picture on the basis of the vector and adopting such vector which minimizes the difference mentioned above.
  • use of the reliability indicating similarity degree H taught by the invention allows the decision as to the optimal vector to be made with higher reliability than the decision based on the inter-picture difference mentioned above.
  • the picture signal processing system includes a temporal band limiter circuit 102, a motion vector detecting apparatus 104, a delay circuit 106 and a processing unit 108.
  • the circuit 102 is implemented in a configuration similar to that of the section 2 shown in Fig. 1.
  • the detecting apparatus 104 includes the section 3, 4 and 5 shown in Fig. 1 and may further include the section 6 or 7.
  • the circuit 106 serves for delaying transmission of the picture signal to the processing unit 106 by a time which corresponds to the time required for the operation of the detecting apparatus 104 and includes a plurality of field memories. To this end, the plural field memories of the circuit 102 may be shared by the delay circuit 106.
  • the processing unit 108 performs predetermined processings on the picture signal in accordance with the motion vector detected by the detecting apparatus 104.
  • the predetermined processings may comprise a high efficiency encoding processing such as motion compensating encoding or the like, estimation of three-dimensional motion of an object, system transformation processing between NTSC system of 52 lines/49.96 fields and PAL or SECAM systems of 625 lines/50 fields or between 2-to-1 interlacing HDTV system of 1125 lines/60 fields and 2-to-1 interlacing system of 525 lines/59.94 fields or the like.
  • a picture signal processing system shown in Fig. 10 differs from the system shown in Fig. 9 only in respect that the circuit 102 is spared. This is because the motion vector can be determined notwithstanding of absence of the circuit 102 in the conventional television cameras which can exhibit the storage effect or the after-image effort.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Image Analysis (AREA)
  • Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)

Claims (18)

  1. Dispositif de détection de vecteur de mouvement, comprenant:
    un moyen formant mémoire (3) servant à stocker à intervalle prédéterminé des signaux d'image introduits de manière renouvelée;
    un moyen (4) servant à générer deux signaux d'image en appliquant une compression de bande dans l'espace à deux parties de signal dudit signal d'image renouvelé stockées dans ledit moyen formant mémoire (3), qui correspondent respectivement à des intervalles de temps différents; et
    un moyen de détection (5) comprenant au moins un étage de détection (22a) servant à détecter un vecteur de mouvement selon un procédé à gradient, sur la base desdits deux signaux d'image,
    dans lequel ledit moyen de détection (5) comprend un moyen (49) servant à générer une très faible valeur de signal d'activation et un moyen (42) servant à ajouter une très faible valeur de signal d'activation ainsi générée à un gradient déterminé d'après un signal d'image de référence qui est représenté par un desdits deux signaux d'image.
  2. Dispositif de détection de vecteur de mouvement selon la revendication 1, comprenant en outre un moyen (2) servant à appliquer une compression de bande dans le temps à des signaux d'image originaux pour ainsi générer lesdits signaux d'image qui sont fournis audit moyen formant mémoire (3).
  3. Dispositif de détection de vecteur de mouvement selon la revendication 1, comprenant en outre:
    une pluralité de sections (2-1 à 2-4) de mémoire servant respectivement à stocker une pluralité de signaux successifs résultant de la division dudit signal d'image original lors dudit intervalle de temps prédéterminé;
    des moyens de pondération (2-11 à 2-15) servant à pondérer, suivant des coefficients préétablis (K1 à K5), ledit signal d'image original et lesdits signaux stockés dans lesdites plusieurs sections (2-1 à 2-4) de mémoire; et
    des moyens de totalisation (2-20) servant à additionner les uns avec les autres lesdits signaux pondérés afin d'obtenir un signal d'image renouvelé à fournir audit moyen formant mémoire (3).
  4. Dispositif de détection de vecteur de mouvement selon la revendication 2 ou 3, dans lequel ledit étage de détection dudit moyen de détection comprend:
    un moyen (34) servant à déterminer des valeurs de gradients dudit signal d'image de référence au niveau de pixels individuels dans un bloc concerné, chacun desdits signaux d'image étant divisé en une pluralité de blocs comprenant chacun une pluralité de pixels, ledit bloc concerné étant un desdits plusieurs blocs pour lesquels ledit vecteur de mouvement est à détecter;
    un moyen (42, 44, 49) servant à déterminer des valeurs absolues en ajoutant ladite très faible valeur auxdites valeurs de gradients;
    un moyen (32) servant à déterminer, suivant les valeurs de gradients, des différences de valeur entre lesdits deux signaux d'image au niveau desdits pixels;
    des moyens de totalisation (38, 46) servant respectivement à additionner lesdites valeurs absolues et lesdites différences pour la totalité desdits pixels dans ledit bloc concerné; et
    un moyen (48) servant à déterminer une valeur d'une composante dudit vecteur de mouvement en divisant une somme desdites différences par une somme desdites valeurs absolues.
  5. Dispositif de détection de vecteur de mouvement selon l'une quelconque des revendications précédentes, comprenant en outre un moyen d'évaluation (7) servant à déterminer entre deux points sur un espace à N dimensions une distance spécifiée par deux ensembles de coefficients obtenues en appliquant une transformation orthogonale de degré N, N étant un entier non inférieur à 2, à une partie du signal d'image de référence dans ledit bloc concerné pour lequel ledit vecteur de mouvement a été détecté et à une partie de signal obtenue en décalant par ledit vecteur de mouvement une partie de signal correspondant à l'autre desdits deux signaux d'image correspondant audit bloc concerné, afin de déterminer ainsi un degré de similitude entre lesdites deux parties de signal d'image pour évaluer la fiabilité dudit vecteur de mouvement.
  6. Dispositif de détection de vecteur de mouvement selon la revendication 5, dans lequel ledit moyen d'évaluation (7) comprend en outre un moyen (70) servant à évaluer la fiabilité dudit vecteur de mouvement avec une valeur obtenue en normalisant la distance entre lesdits deux points par des distances de l'origine dudit espace à N dimensions auxdits deux points.
  7. Dispositif de détection de vecteur de mouvement selon l'une quelconque des revendications précédentes, dans lequel ledit moyen de détection comprend:
    un moyen de fourniture (20) de vecteur de déplacement initial;
    un moyen de détection (22a) servant à détecter un vecteur de déplacement sur la base du vecteur de déplacement initial en fonction desdits deux signaux d'image et de ladite très faible valeur; et
    un moyen servant à détecter le vecteur de mouvement à partir dudit vecteur de déplacement initial et dudit vecteur de déplacement détecté.
  8. Dispositif de détection de vecteur de mouvement selon l'une quelconque des revendications 1 à 6, dans lequel ledit moyen de détection comprend:
    un moyen (20) servant à fournir un vecteur de déplacement initial; et
    une pluralité (22a, 22b, 22c) d'étages de détection servant chacun à détecter son propre vecteur en fonction d'un vecteur de déplacement fourni depuis un étage précédent, desdits deux signaux d'image et de la très faible valeur, pour déterminer de la sorte son propre vecteur de déplacement fourni depuis l'étage précédent et ledit propre vecteur détecté, ledit propre vecteur de déplacement étant délivré à un étage suivant, dans lequel un étage initial (22a) desdits plusieurs étages de détection (22a, 22b, 22c) est fourni avec le vecteur de déplacement initial par ledit moyen d'alimentation, tandis qu'un étage final (22c) de ladite pluralité d'étages de détection (22a, 22b, 22c) délivre son propre vecteur de déplacement en tant que dit vecteur de mouvement.
  9. Dispositif de détection de vecteur de mouvement selon la revendication 7 ou 8, comprenant en outre un moyen (56) de détection de changement de scène pour détecter un changement de scène en fonction d'un certain nombre des vecteurs de déplacement dont la valeur est supérieure à une première valeur prédéterminée.
  10. Système de traitement de signal d'image, comprenant:
    une section (104) de détection de vecteur de mouvement servant à détecter un vecteur de mouvement selon un procédé à gradient sur la base de deux signaux d'images renouvelés associés à deux des parties de signal d'image d'un signal d'image original, dans lequel ladite section de détection (104) comprend un moyen (49) servant à générer une très faible valeur de signal d'activation et un moyen (42) servant à additionner une très faible valeur de signal d'activation ainsi générée avec un gradient déterminé d'après un desdits deux signaux d'image renouvelés;
    un moyen à retard (106) servant à retarder ledit signal d'image original dans une mesure correspondant à une pluralité de laps de temps prédéterminés; et
    un moyen de traitement (109) servant à effectuer un traitement prédéterminé sur ledit signal d'image original retardé par ledit moyen à retard (106) sur la base du vecteur de mouvement détecté.
  11. Système de traitement de signal d'image selon la revendication 10, dans lequel ladite section de détection de vecteur de mouvement comprend:
    un moyen formant mémoire (3) servant à stocker lesdits deux signaux d'image renouvelés;
    un moyen (4) servant à générer deux signaux traités en appliquant une compression de bande dans l'espace respectivement auxdits deux signaux d'image renouvelés stockés dans ledit moyen formant mémoire (3); et
    un moyen de détection (5) servant à détecter le vecteur de mouvement suivant le procédé à gradient sur la base desdits deux signaux traités et de la très faible valeur à ajouter au gradient déterminé d'après un signal de référence représenté par un desdits deux signaux traités.
  12. Application de traitement de signal d'image selon la revendication 11, dans laquelle ledit moyen de détection (5) comprend:
    un moyen (34) servant à déterminer des valeurs de gradients dudit signal de référence au niveau de pixels individuels dans un bloc concerné, chacun desdits signaux d'image étant divisé en une pluralité de blocs contenant chacun une pluralité de pixels, ledit bloc concerné étant un desdits plusieurs blocs pour lesquels ledit vecteur de mouvement est à détecter;
    un moyen (42, 44, 49) servant à déterminer des valeurs absolues en ajoutant ladite très faible valeur auxdites valeurs de gradients;
    un moyen (32) servant à déterminer, en fonction des valeurs de gradients, des différences de valeur entre lesdits deux signaux traités au niveau desdits pixels;
    des moyens de totalisation (38, 46) servant respectivement à totaliser lesdites valeurs absolues et lesdites différences pour la totalité desdits pixels dans ledit bloc concerné; et
    un moyen (48) servant à déterminer une valeur d'une composante dudit vecteur de mouvement en divisant une somme desdites différences par une somme desdites valeurs absolues.
  13. Système de traitement de signal d'image selon la revendication 11, dans lequel ledit moyen de détection (5) comprend:
    un moyen (20) servant à fournir un vecteur de déplacement initial; et une pluralité (22a, 22b, 22c) d'étages de détection chacun pour détecter son propre vecteur en fonction d'un vecteur de déplacement fourni depuis un étage précédent, desdits deux signaux traités et de ladite très faible valeur, pour déterminer de la sorte son propre vecteur de déplacement d'après le vecteur de déplacement fourni depuis l'étage précédent et ledit propre vecteur détecté, ledit propre vecteur de déplacement étant délivré à un étage suivant parmi lesdits étages de détection, dans lequel un étage initial (22a) desdits plusieurs étages de détection (22a, 22b, 22c) reçoit dudit moyen d'alimentation le vecteur de déplacement initial, tandis qu'un étage final (22c) desdits plusieurs étages de détection (22a, 22b, 22c) délivre sont propre vecteur de déplacement en tant que dit vecteur de mouvement.
  14. Système de traitement de signal d'image selon la revendication 13, dans lequel ladite section de détection de vecteur de mouvement comprend en outre un moyen (36) de détection de changement de scène (36) servant à prendre une décision, bloc par bloc, pour établir si ledit vecteur de déplacement fourni depuis un étage prédéterminé parmi lesdits plusieurs étages de détection est ou n'est pas supérieur à une première valeur prédéterminée, pour détecter de la sorte un changement de scène en fonction d'un certain nombre des vecteurs de déplacement ayant une valeur supérieure à ladite première valeur prédéterminée.
  15. Système de traitement de signal d'image selon l'une quelconque des revendications 11 à 14, comprenant en outre un moyen (102) servant à appliquer un effet de stockage et/ou un effet de traínage, pendant ledit laps de temps prédéterminé, audit signal retardé par ledit moyen à retard, pour générer de ce fait lesdits deux signaux d'image renouvelés à fournir à ladite section de détection.
  16. Système de traitement de signal d'image selon l'une quelconque des revendications 11 à 15, dans lequel ladite section de détection de vecteur de mouvement comprend en outre un moyen d'évaluation (7) servant à déterminer une distance entre deux points sur un espace à N dimensions spécifié par deux ensembles de coefficients obtenus en appliquant une transformation orthogonale de degré N, N étant un entier non inférieur à 2, à une partie d'un desdits deux signaux d'image renouvelés correspondant au bloc concerné pour lequel ledit vecteur de mouvement a été détecté et une partie de l'autre desdits deux signaux d'image correspondant audit bloc concerné, pour évaluer de la sorte la fiabilité dudit vecteur de mouvement sur la base de ladite distance.
  17. Procédé de détection de vecteur de mouvement à partir de signaux d'image, comprenant les étapes consistant à:
    fournir un vecteur de déplacement initial;
    détecter un vecteur dans un étage donné, suivant un procédé à gradient, sur la base du vecteur de déplacement et de deux signaux d'image fournis, consistant à générer une très faible valeur de signal d'activation et à ajouter une très faible valeur de signal d'activation ainsi générée à un gradient déterminé à partir d'un signal d'image de référence représenté par un desdits deux signaux d'image;
    déterminer un vecteur de déplacement dans l'étage donné à partir du vecteur de déplacement et du vecteur détecté fourni et délivrer à un étage suivant le vecteur de déplacement de l'étage donné; et
    répéter un certain nombre de fois, correspondant à un nombre d'étages de détection, ladite étape de détection et ladite étape de détermination de vecteur de déplacement, un étage initial parmi lesdits étages recevant ledit vecteur de déplacement initial en tant que dit vecteur de déplacement, tandis qu'un étage final parmi lesdits étages délivre le vecteur de déplacement en tant que vecteur de mouvement.
  18. Procédé selon la revendication 17, dans lequel ladite étape de détection comprend les étapes consistant à:
    déterminer des valeurs de gradients d'un des deux signaux d'image au niveau de pixels individuels dans un bloc concerné, chacun desdits signaux d'image étant divisé en une pluralité de blocs comportant chacun une pluralité de pixels, ledit bloc concerné étant un desdits plusieurs blocs pour lequel ledit vecteur de mouvement est à détecter;
    ajouter ladite très faible valeur aux valeurs de gradients, puis déterminer des valeurs absolues de celles-ci;
    déterminer des différences de valeurs respectivement entre les deux signaux d'image au niveau desdits pixels, lesdites différences étant déterminées dans la direction déterminée par une composante directionnelle du vecteur de mouvement à détecter;
    totaliser respectivement les valeurs absolues et les différences pour tous les pixels dans le bloc concerné; et
    déterminer la composante directionnelle dudit vecteur de mouvement en divisant une valeur de somme des différences par une valeur de somme des valeurs absolues.
EP92910679A 1991-05-23 1992-05-22 Procede pour detecter un vecteur de mouvement et appareil a cet effet, et systeme pour traiter un signal video a l'aide de cet appareil Expired - Lifetime EP0540762B1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
EP97110945A EP0806866B1 (fr) 1991-05-23 1992-05-22 Dispositif d'évaluation et méthode utilisable dans un appareil de détection de vecteurs de mouvement

Applications Claiming Priority (13)

Application Number Priority Date Filing Date Title
JP14662591 1991-05-23
JP146624/91 1991-05-23
JP146625/91 1991-05-23
JP14662491A JP2898787B2 (ja) 1991-05-23 1991-05-23 動きベクトル検出装置
JP14662491 1991-05-23
JP3146625A JP3067275B2 (ja) 1991-05-23 1991-05-23 シーンチェンジ検出装置
JP16625491 1991-06-12
JP166254/91 1991-06-12
JP16625491A JP2989325B2 (ja) 1991-06-12 1991-06-12 動きベクトル信頼性の判定装置
JP20117091A JP2898798B2 (ja) 1991-07-17 1991-07-17 勾配法による動きベクトル検出装置
JP201170/91 1991-07-17
JP20117091 1991-07-17
PCT/JP1992/000659 WO1992021210A1 (fr) 1991-05-23 1992-05-22 Procede pour detecter un vecteur en mouvement et appareil a cet effet, et systeme pour traiter un signal video a l'aide de cet appareil

Related Child Applications (1)

Application Number Title Priority Date Filing Date
EP97110945A Division EP0806866B1 (fr) 1991-05-23 1992-05-22 Dispositif d'évaluation et méthode utilisable dans un appareil de détection de vecteurs de mouvement

Publications (3)

Publication Number Publication Date
EP0540762A1 EP0540762A1 (fr) 1993-05-12
EP0540762A4 EP0540762A4 (en) 1994-08-17
EP0540762B1 true EP0540762B1 (fr) 1999-08-04

Family

ID=27472727

Family Applications (2)

Application Number Title Priority Date Filing Date
EP92910679A Expired - Lifetime EP0540762B1 (fr) 1991-05-23 1992-05-22 Procede pour detecter un vecteur de mouvement et appareil a cet effet, et systeme pour traiter un signal video a l'aide de cet appareil
EP97110945A Expired - Lifetime EP0806866B1 (fr) 1991-05-23 1992-05-22 Dispositif d'évaluation et méthode utilisable dans un appareil de détection de vecteurs de mouvement

Family Applications After (1)

Application Number Title Priority Date Filing Date
EP97110945A Expired - Lifetime EP0806866B1 (fr) 1991-05-23 1992-05-22 Dispositif d'évaluation et méthode utilisable dans un appareil de détection de vecteurs de mouvement

Country Status (4)

Country Link
US (1) US5436674A (fr)
EP (2) EP0540762B1 (fr)
DE (2) DE69230115T2 (fr)
WO (1) WO1992021210A1 (fr)

Families Citing this family (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
USRE39279E1 (en) * 1991-11-08 2006-09-12 Matsushita Electric Industrial Co., Ltd. Method for determining motion compensation
US5369449A (en) * 1991-11-08 1994-11-29 Matsushita Electric Industrial Co., Ltd. Method for predicting move compensation
USRE39276E1 (en) * 1991-11-08 2006-09-12 Matsushita Electric Industrial Co., Ltd. Method for determining motion compensation
DE4311972A1 (de) * 1993-04-10 1994-10-13 Bosch Gmbh Robert Verfahren zur Detektion von Änderungen in Bewegtbildern
JPH07135663A (ja) * 1993-09-17 1995-05-23 Oki Electric Ind Co Ltd 動きベクトル検出方法および動きベクトル検出装置
US5682438A (en) * 1994-06-21 1997-10-28 Nippon Steel Corporation Method and apparatus for determining a motion vector
US5889891A (en) * 1995-11-21 1999-03-30 Regents Of The University Of California Universal codebook vector quantization with constrained storage
DE19626108A1 (de) * 1996-06-28 1998-01-08 Thomson Brandt Gmbh Verfahren und Vorrichtung zur Codierung von digitalen Bildsignalen
JPH10336668A (ja) * 1997-06-02 1998-12-18 Sharp Corp 動きベクトル検出装置
JP3409834B2 (ja) * 1997-07-10 2003-05-26 ソニー株式会社 画像処理装置および画像処理方法、並びに記録媒体
US6618507B1 (en) 1999-01-25 2003-09-09 Mitsubishi Electric Research Laboratories, Inc Methods of feature extraction of video sequences
US6625216B1 (en) * 1999-01-27 2003-09-23 Matsushita Electic Industrial Co., Ltd. Motion estimation using orthogonal transform-domain block matching
EP1065877A1 (fr) * 1999-07-01 2001-01-03 Mitsubishi Denki Kabushiki Kaisha Identification de segments de trains binaires vidéo comprimés
EP1104197A3 (fr) * 1999-11-23 2003-06-04 Texas Instruments Incorporated Compensation de mouvement
JP4063486B2 (ja) * 2000-09-06 2008-03-19 日本電気株式会社 画像検索装置及び方法並びに画像検索プログラムを記録した記憶媒体
ES2526340T3 (es) 2001-07-10 2015-01-09 Entropic Communications, Inc. Unidad y método para la estimación del movimiento, y aparato de procesamiento de imagen provisto con tal unidad de estimación de movimiento
EP1387340A1 (fr) * 2002-07-30 2004-02-04 Deutsche Thomson-Brandt Gmbh Procédé et dispositif pour le traitement des données d'image vidéo pour un dispositif d'affichage
US20050084135A1 (en) * 2003-10-17 2005-04-21 Mei Chen Method and system for estimating displacement in a pair of images
EP2103125B1 (fr) * 2007-01-08 2016-08-17 SK Telecom Co., Ltd. Système et procédé assurant une synchronisation d'un contenu d'une émission avec une information supplémentaire
DE102007027642A1 (de) * 2007-06-15 2008-12-18 Micronas Gmbh Verfahren zur Bearbeitung einer Bildfolge mit aufeinanderfolgenden Videobildern zur Verbesserung der räumlichen Auflösung
US8326042B2 (en) * 2007-06-18 2012-12-04 Sony (China) Limited Video shot change detection based on color features, object features, and reliable motion information
US7990476B2 (en) * 2007-09-19 2011-08-02 Samsung Electronics Co., Ltd. System and method for detecting visual occlusion based on motion vector density
US20090268097A1 (en) * 2008-04-28 2009-10-29 Siou-Shen Lin Scene change detection method and related apparatus according to summation results of block matching costs associated with at least two frames
FI20095273A0 (fi) * 2009-03-17 2009-03-17 On2 Technologies Finland Oy Digitaalinen videokoodaus
EP2405661B1 (fr) 2010-07-06 2017-03-22 Google, Inc. Transmission vidéo résistante à la perte utilisant deux décodeurs
US9014265B1 (en) 2011-12-29 2015-04-21 Google Inc. Video coding using edge detection and block partitioning for intra prediction
US9210424B1 (en) 2013-02-28 2015-12-08 Google Inc. Adaptive prediction block size in video coding
US9313493B1 (en) 2013-06-27 2016-04-12 Google Inc. Advanced motion estimation
US9807416B2 (en) 2015-09-21 2017-10-31 Google Inc. Low-latency two-pass video coding
CN106131592A (zh) * 2016-07-18 2016-11-16 乐视控股（北京）有限公司 一种电视频道的确定方法和装置

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2549329B1 (fr) * 1983-07-13 1987-01-16 Thomson Csf Procede et dispositif de detection de points en mouvement dans une image de television pour systemes de television numerique de compression de debit a rafraichissement conditionnel
CA1246734A (fr) * 1983-09-08 1988-12-13 Nec Corp Dispositif de detection de mouvements dans les signaux de television
JPS60158786A (ja) * 1984-01-30 1985-08-20 Kokusai Denshin Denwa Co Ltd <Kdd> 画像動き量検出方式
JPS60189388A (ja) * 1984-03-09 1985-09-26 Fujitsu Ltd 動き補償符号化装置
CA1277416C (fr) * 1984-08-13 1990-12-04 Akihiro Furukawa Appareil de codage predictif inter-trame pour signaux video
WO1987005769A1 (fr) * 1986-03-19 1987-09-24 British Broadcasting Corporation Mesurage du mouvement d'images tv
JPS62219787A (ja) * 1986-03-20 1987-09-28 Hitachi Ltd 動き検出測定方式
JP2768669B2 (ja) * 1987-01-23 1998-06-25 株式会社日立製作所 Ｔｖ信号の動き補償フレーム間符号化装置
CA1333502C (fr) * 1987-08-28 1994-12-13 Michael Douglas Carr Codage de signaux
DE3809249A1 (de) * 1988-03-18 1989-09-28 Thomson Brandt Gmbh Bewegungsdetektor
CA1322241C (fr) * 1988-07-22 1993-09-14 Hisafumi Motoe Circuit de detection de mouvement
JPH0683435B2 (ja) * 1989-02-03 1994-10-19 三洋電機株式会社 サブサンプル映像信号復調装置
JPH0810936B2 (ja) * 1989-03-31 1996-01-31 松下電器産業株式会社 動きベクトル検出装置
FR2648590B1 (fr) * 1989-06-20 1991-08-30 Thomson Consumer Electronics Procede et dispositif d'estimation de mouvement dans une sequence d'images animees
US5083206A (en) * 1990-03-19 1992-01-21 At&T Bell Laboratories High definition television arrangement including noise immunity means
KR930010359B1 (ko) * 1991-06-27 1993-10-16 삼성전자 주식회사 영상기록장치
US5237413A (en) * 1991-11-19 1993-08-17 Scientific-Atlanta, Inc. Motion filter for digital television system

Also Published As

Publication number Publication date
EP0806866A2 (fr) 1997-11-12
US5436674A (en) 1995-07-25
DE69230115D1 (de) 1999-11-11
EP0540762A4 (en) 1994-08-17
DE69229705T2 (de) 1999-12-23
DE69229705D1 (de) 1999-09-09
EP0540762A1 (fr) 1993-05-12
WO1992021210A1 (fr) 1992-11-26
EP0806866A3 (fr) 1997-11-19
EP0806866B1 (fr) 1999-10-06
DE69230115T2 (de) 2000-04-20

Similar Documents

Publication Publication Date Title
EP0540762B1 (fr) Procede pour detecter un vecteur de mouvement et appareil a cet effet, et systeme pour traiter un signal video a l&#39;aide de cet appareil
US7336316B2 (en) De-interlacing of video data
US5920657A (en) Method of creating a high resolution still image using a plurality of images and apparatus for practice of the method
US6940557B2 (en) Adaptive interlace-to-progressive scan conversion algorithm
KR100335862B1 (ko) 에지 상관을 사용하여 인터레이스 비디오를 프로그레시브비디오로 변환하는 시스템
US8243194B2 (en) Method and apparatus for frame interpolation
US20020159527A1 (en) Reducing halo-like effects in motion-compensated interpolation
WO2007051993A1 (fr) Detection de mouvement video
EP1444837B1 (fr) Procede et dispositif de detection d&#39;occlusions
US20060188154A1 (en) Digital video signal processing apparatus and method for compensating a chrominance signal
US7679676B2 (en) Spatial signal conversion
EP1863283B1 (fr) Méthode et dispositif d&#39;interpolation d&#39;image
KR20060135667A (ko) 이미지 포맷 변환
Jung et al. An effective de-interlacing technique using two types of motion information
AU2004200237B2 (en) Image processing apparatus with frame-rate conversion and method thereof
EP1636987B1 (fr) Conversion spatiale de signaux
EP1574055A1 (fr) Reconnaissance de films et de video apparaissant en parallele sur des champs de television
EP0648046A2 (fr) Méthode et dispositif pour interpoler avec compensation de mouvement des champs ou des trames intermédiaires
US20060038918A1 (en) Unit for and method of image conversion
Lachine et al. Edge adaptive intra field de-interlacing of video images
Weiss An improved algorithm for deinterlacing video streams
Jung et al. Deinterlacing method based on edge direction refinement using weighted maximum frequent filter

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19921216

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): DE FR GB

A4 Supplementary search report drawn up and despatched
AK Designated contracting states

Kind code of ref document: A4

Designated state(s): DE FR GB

17Q First examination report despatched

Effective date: 19961028

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE FR GB

DX Miscellaneous (deleted)
ET Fr: translation filed
REF Corresponds to:

Ref document number: 69229705

Country of ref document: DE

Date of ref document: 19990909

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed
REG Reference to a national code

Ref country code: GB

Ref legal event code: IF02

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20110523

Year of fee payment: 20

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: GB

Payment date: 20110518

Year of fee payment: 20

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20110518

Year of fee payment: 20

REG Reference to a national code

Ref country code: DE

Ref legal event code: R071

Ref document number: 69229705

Country of ref document: DE

REG Reference to a national code

Ref country code: DE

Ref legal event code: R071

Ref document number: 69229705

Country of ref document: DE

REG Reference to a national code

Ref country code: GB

Ref legal event code: PE20

Expiry date: 20120521

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF EXPIRATION OF PROTECTION

Effective date: 20120523

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: GB

Free format text: LAPSE BECAUSE OF EXPIRATION OF PROTECTION

Effective date: 20120521